Virtual Computing Environment for Future Combat Systems

1
Virtual Computing Environment for Future Combat
Systems
2
Participating Teams and Team Members

• Participating Teams
• Enabling Technologies in Support of NCC
• Computational Structural Mechanics and
  material-by-design
• Chem-bio
• Team Members
• Robert Baffeur
• Baoquan Chen
• Ravi Janardan
• Shashi Shekhar
• Kumar Tamma
• Jon Weissman

3
Motivating Example

• Examples
• Chem-Bio portfolio project (Dr. Alibadi)
• Scenario managing a (say chem-bio) attack
• Components of the system
• Gathering initial conditions
• Weather data from NWS or JSU
• Terrain maps (State of federal Govt.)
• Building geometry (City Govt.)
• Plume simulation using supercomputers
• Visualizing results map, 3D graphics
• Response planning
• Current effort is very large because of
• Lack of large-scale reusable components
• Lack of integrations services
• Autonomous components
• Distribution location of data, computers
• Heterogeneity e.g. data formats
• Service model interfaces are low-level

Weather, Terrain, Base map
Plume Modeling
Demographics, Transportation
( Images from www.fortune.com )
4
Problem Definition

• Other Examples
• Projectile-Target interaction portfolio
• Precision targeting of missiles
• Other Logistics of troop/equipment movement
• Deadlines, congstion, price,
• Issue Mega-programming Wiederhold 98
• Software size is growing
• Shift in Programming tasks
• Larger systems
• Reuse of well-defined components
• Need more integration effort
• Current tools are inadequate
• Few reusable components
• Little support for component integration

Integration
Coding
Small
Large
Medium
5
Technical Goals

• Develop a VCE
• To support rapid development of
• for multi-disciplinary large-scale software
  development
• To increase the lethality and survivability of
  Army FCS
• What is a VCE?
• Tools to integrate distributed heterogeneous
  autonomous components
• Data format interchange (ICE, J. Clarke)
• Resource identification, scheduling (J. Weissman)
• A collection of reusable components
• Large-scale autonomous
• With well-defined services
• Example components
• Visualization (B. Chen)
• Routing (S. Shekhar)
• Prototyping (R. Janardan)
• Virtual HPC design(K. Tamma)

6
VCE - System Architecture
 
                                         
Fluid Dynamic Simulation   Plume simulation of
toxic agents  
GUI   User interface to initialize seat occupancy
and toxic location
V D G D I C E
Database   Graphic objects e.g. 3D Model of
building GIS layer e.g. sections, exits,
capacity
Visualization Engine   Animation of plume
evolution and people movements in the building
User
Physical Prototyping   Create 3D physical
prototype of terrain
HPGIS   Route Planning for evacuation with
capacity constraints and toxic constraints
7
Services by Components

• HPGIS
• Develop algorithms for computing evacuation
  route plans
• Visualization Engine
• Develop graphic visualization tools to display
  evolution of plume and moving of people in the
  building
• Fluid Dynamic Simulation
• Predict and simulate the evolution of plume in
  the building
• Physical Prototyping
• Efficient algorithms and software to create
  physical prototypes of terrain
• Virtual Data Grid
• Provide an interface to underlying Grid
  infrastructure (Globus) to support remote
  execution of components, provide a VDG interface
  to ICE and Globus
• ICE/DICE
• Test ICE/DICE system that has interface for
  exchanging data with other applications for
  real-time visualization.

8
Main Activities

• Tools for integration of components
• Installed ICE system and Globus on Army Center
  machines
• Virtual data grid design and evaluation
• to automate the scheduling and execution of
  components
• Reusable Components
• 3D visualization of Atlanta plume evolution data
  from Prof. Alibadi
• GIS Geo-registration of Atlanta buildings,
  Addition of roads, vegetation
• Dynamic Routing algorithms for Atlanta roadmap in
  presence of plumes
• Prototyping algorithms to prototype terrain

9
Status

• ICE system installation
• QingSong Lu attended ICE workshop in June
• We have an installation of ICE
• Acquisition of 3D Models
• Acquired St. Paul downtown model, Atlanta
  downtown model with plume simulation
• Tested St. Paul model with graphics rendering
  programs
• Acquisition of Visualization Environments
• Reviewed visualization equipment at ARL, Iowa
  State U, MechDyne, SGI, etc.
• Working on getting a configuration to start with
  and grow with
• Development of Visualization Software
• Tested St. Paul model with graphics rendering
  software
• Working with Prof. Candler to visualize
  fluid-flows over urban terrain
• HPGIS Route Planning Algorithms
• Surveyed literature on routing algorithms and
  evacuation planning
• Formulated the problem and developed preliminary
  approaches

10
Rapid Protoyping Project Goal
Develop scalable HPC techniques that allow the
effective deployment of RPP in Battlefield
Visualization.
Create, on demand, physical scale models of enemy
terrains and assets to help mission planners and
field commanders develop and evaluate different
combat strategies. Digital data generated from
satellite images, airborne laser scanners, etc.
RPP can also be used in Signature Modeling to
create detailed physical scale models of complex
weapons systems (e.g., tanks, UAVs) that can then
be tested for their signature mitigation
capabilities (e.g., radar evasion).
11
Work done in Year 1
Investigated efficient methods for identifying
structural properties of battlefield
topographies.
12
Terrain Recognition Algorithm
Designed very efficient algorithm to decide if a
cross-section of a topography is a 2D terrain in
some direction.
Computational effort scales just linearly with
size of cross-section. Result based on geometric
properties that depend only on edge orientations,
not on global connectivity of polygon.
Yields a fast, first-cut test for determining
if given topography is a 3D terrain.
13
Terrain Decomposition Algorithm
Designed efficient algorithm to decompose a 2D
near-terrain into two terrains.
Computational effort scales roughly from linear
to quadratic. Searches for decomposition via a
rotational sweep on convex hull.
Allows for fast building of the two pieces in
parallel.
14
Innovative HPC design and Analysis Approaches for
Flexible Multi-body Dynamics K Tamma

• There is a need to design/develop and implement a
  general purpose flexible multibody dynamics code
  which possesses the following capabilities for
  general applications of Army interest
• Efficiently integrate Index-3 stiff DAE,
  preserving the order of accuracy in
  primary/algebraic variables by using either
• Index-3 formulation.
• Stabilization
• Constraint preservation
• Prevent order reduction
• Be able to represent large deformations/strain
  accurately as in the case of inertial coordinate
  formulation and finite deformation dynamics.
• Invariant conserving algorithms for long term
  rigid/flexible multibody dynamics.
• Optimal/Intelligent dissipation in order to damp
  out high frequency oscillations and smart
  integrators.
• Be able to model contact physics at joints.
• HPC simulations and parallel environments

15
Multibody systems

• Multibody system is defined to be a collection of
  bodies, which are kinematically constrained due
  to different types of joints. Each body may
  undergo large translations rotations.

Joint
Body 1
Body 2
Force Element
Damper
Actuator
Body 3
Body n
Multibody System
Some Common Joints
Fundamental Elements of Multibody System
Rigid Body
16
Progress
Andrews seven body squeezer mechanism New
developments prevented order reduction
2
1
2
New
Existing
Accuracy Velocity
Simulation of a spin top motion using energy
conserving scheme
17
Next Few Talks on VCE Portfolio

• HPGIS Next talk by S. Shekhar
• Visualization Talk by B. Chen
• Virtual Data Grid Talk by J. Weissman